Abstract
Habitat loss and fragmentation are the leading causes of species range contraction and extirpation, worldwide. Factors that predict sensitivity to fragmentation include high trophic level, large body size, and extensive spatial requirements. Pumas (Puma concolor) exemplify these qualities, making them particularly susceptible to fragmentation and subsequent reductions in demographic connectivity. The chaparral-dominated ecosystems surrounding the greater San Francisco Bay Area encompass over 10,000 km(2) of suitable puma habitat, but inland waterways, croplands, urban land uses, and extensive transportation infrastructure have resulted in widespread habitat fragmentation. Pumas in this region now exist as a metapopulation marked by loss of genetic diversity, collisions with vehicles, and extensive human-puma conflict. Given these trends, we conducted a photo survey from 2017 to 2021 across 19 patches of predicted habitat and compiled a dataset of >6584 puma images. We used a logistic regression analytical framework to evaluate the hypothesis that puma patch occupancy would exhibit a threshold response explained by patch size, isolation, and habitat quality. Contrary to predictions, only variables related to patch size demonstrated any power to explain occupancy. On average, occupied patches were 18× larger than those where they were not detected (825 ± 1238 vs. 46 ± 101 km(2)). Although we observed pumas in patches as small as 1 km(2), logistic regression models indicated a threshold occupancy probability between 300 and 400 km(2), which is remarkably close to the mean male puma home range size in coastal California (~381 km(2)). Puma populations dependent on habitats below this value may be susceptible to inbreeding depression and human-wildlife conflict, and therefore vulnerable to extirpation. For species conservation, we suggest conflicts might be ameliorated by identifying the largest, isolated patches for public education campaigns with respect to management of domestic animals, and remaining connective parcels be identified, mapped, and prioritized for targeted mitigation.